https://bugs.documentfoundation.org/show_bug.cgi?id=90235
Matthew Francis <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Keywords| |bisected, regression CC| |[email protected] Summary|It's a "sum" function bug? |Loss of precision in SUM | |function Whiteboard| |bibisected OS|Windows (All) |All --- Comment #5 from Matthew Francis <[email protected]> --- Forcing recalculation to override the cached value in the file gives a result of 0.00 until source-hash-61db96daa87754af24355d7ac94ee0305f22ff87 in bibisect 43all (* recalculation by editing the cell - F9 doesn't seem to affect this for some reason) At a quick scan this is most likely a result of the below commit commit 9d207f6fc479bf96db9fcde214f15f2ffb045001 Author: Kohei Yoshida <[email protected]> Date: Tue Jul 2 00:51:10 2013 -0400 Re-implement SUM function to make use of new cell storage. And fix a bug in my new COUNT function, where I forgot to tally numeric formula cells. Change-Id: I52d26be3e48f646f656821066e23594d52f78c6d and/or this one which followed it commit d61233a388284383a939e2f26a0e25b9ad243898 Author: Kohei Yoshida <[email protected]> Date: Tue Jul 2 16:02:59 2013 -0400 Move this code from the column code back into the interpreter code. Change-Id: I7830cdf3f09ed7b6ae6221212bfb84abcdeac523 Still not sure if this is a genuine bug or just an expected consequence of floating point arithmetic. However, as it used to give an exact result, let's call this a regression until a Calc developer decides otherwise. -- You are receiving this mail because: You are the assignee for the bug.
_______________________________________________ Libreoffice-bugs mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/libreoffice-bugs
